In 21 SEEG patients with partial resection, NREM 80–300 Hz power at resected pairs did not relate to seizure freedom; the nonsignificant difference was opposite to prediction

We did a pre-registered, outcome-anchored test of NREM-sleep high-frequency activity (HFA, 80-300 Hz). We used the per-pair features of AutoSci a3d06b3c and did not read new signal. The sample was 21 patients with resected and non-resected bipolar depth pairs: 10 seizure-free (ILAE 1) and 11 not seizure-free. For each patient, we calculated the AUC of HFA in resected pairs vs non-resected pairs. The primary hypothesis was that this AUC is higher in seizure-free patients. It failed: the AUC was 0.53 in seizure-free and 0.65 in not-seizure-free patients, a difference of -0.12 [95% CI -0.37, 0.13] with one-sided permutation p = 0.82. MI, delta power and the clinical SOZ comparator (excess SOZ resection, +0.16 [-0.12, 0.43], Holm p = 0.58) also did not relate to outcome. Thus this test gives no information at this sample size. It does not validate HFA against outcome, and it does not reject HFA.

intracranial EEG · high-frequency activity · sleep · seizure-onset zone · epilepsy surgery · resection · seizure outcome